Research for Predicting Road Skid Resistance Based on Machine Vision

The texture of the pavement surface and its ability to resist the polishing effect of traffic is of prime importance in providing skidding resistance. Skid resistance is a measure of the resistance of pavement surface to sliding or skidding of the vehicle. Road surface texture is essential for pavement skid-resistance. However, its measurement is only possible in the laboratory on cores taken from trafficked roads, and is time consuming. Now it is necessary to develop new methods of monitoring road on-site efficiently. In this dissertation, a system has been developed for measuring the road texture, form and angularity of aggregate based on image analysis, which consists of a line-laser, CCD and a computer. The line-laser can scan a field of 100mm for each time. The form and angularity of aggregate and microtexture of pavement surface is measured by analyzing image of laser. And the character of road microtexture can be recognized by pattern recognition. Thus the methods which were used to measure the shape characteristics of aggregates are unified. Basing on the principles set for the sand patch method, the mathematic model and hardware designing were given. A comparative study was performed to find the British Pendulum Skid Resistance Number for a number of mixes. The experimental results show that the vision measurement system can efficiently improve the testing speed and the accuracy on-site.